# Descriptive and Inferential Statistics

Descriptive Statistics

Inferential Statistics

- Sample vs. Population

### Descriptive Statistics

Descriptive statistics are numbers which are used to describe or summarize your data e.g., average life expectancy in Brazil, per capita income in US etc.

Data (which is a plural indeed!) might have come from Surveys, Experiments or historical data collected in a process.

As the name suggests Descriptive statistics just summarizes the data at hand it does talk about the sample we are studying and we do not infer anything for the population.

When we try to infer the behavior for a larger group that becomes inferential statistics.

For now we will focus on understanding descriptive statistics.The table below shows the literacy rates across various countries, and give an insight how the countries are performing in terms of literacy.

Country | Adult Literacy Rate | Youth Literacy Rate 15-24 age |
---|---|---|

China | 95.1% (2010) | 99.7% (2015) |

Sri Lanka | 92.6% (2015) | 98.8% (2015) |

Myanmar | 89.9% (2007) | 96.3% (2015) |

World Average | 84% (2010) | 89.6% (2010) |

India | 74.04% (2011) | 90.2% (2015) |

Nepal | 55.5% (2007) | 86.9%(2015) |

Pakistan | 50.2% (2007) | 74.8% (2015) |

Bangladesh | 53.5% (2007) | 83.2% (2015) |

(source: wikipedia)

It is a very proud fact tat we are moving towards 100% literacy rate.

For more descriptive analytics you can see the following table which talks about Sex ratio in different states in India. Descriptive statistics is very insightful, e.g., it is clear from the table below that ‘Kerala’ and “Puducherry” have very good Sex Ratio but in the Children it is decreasing which is an alarming situation for us.

2011 Census | 2001 Census | ||||

Rank |
State |
Sex Ratio |
Child Sex ratio |
Sex Ratio |
Child Sex ratio |

1 | Kerala | 1084 | 964 | 1058 | 960 |

2 | Puducherry | 1037 | 967 | 1001 | 967 |

3 | Tamil Nadu | 996 | 943 | 987 | 942 |

4 | Andhra Pradesh | 993 | 939 | 978 | 961 |

5 | Chhattisgarh | 991 | 969 | 989 | 975 |

6 | Meghalaya | 989 | 970 | 972 | 973 |

7 | Manipur | 985 | 930 | 974 | 957 |

8 | Orissa | 979 | 941 | 972 | 953 |

9 | Mizoram | 976 | 970 | 935 | 964 |

10 | Goa | 973 | 942 | 961 | 938 |

11 | Karnataka | 973 | 948 | 965 | 946 |

12 | Himachal Pradesh | 972 | 909 | 968 | 896 |

(Source: Wikipedia)

Descriptive analytics are gateway to understand your data – It is advisable to run descriptive analytics whenever you get a new data to work on. Descriptive analytics motivates us to do further deep dives in data. For example, in the Literacy Data – We would like to further study about the differences in China and Pakistan which causes such a huge difference in literacy rates.

### Inferential Analytics

Inferential analytics is the branch of analytics where we estimate the population behavior from a sample. If you have to understand the brand affinity for your company, it is not possible to go and ask each and every individual on earth, rather a small subset of population (i.e., sample) is selected to study the behavior.

#### Sample Vs. Population

As discussed above, sample is a small portion or subset of the entire population which is a close representative of the population. By close representation of population, we mean that the descriptive statistics for sample and population should be similar. E.g., if population has 50% males and 50% females, it shouldn’t be the case that your sample has 90% males.

There are many techniques in which a sample can be selected. E.g., Simple Random Sampling, Stratified Sampling etc.

Sampling methods will be discussed in later articles.

#### analyticsfreak

#### Latest posts by analyticsfreak (see all)

- Few interesting questions related to correlation - July 22, 2016
- How to make a reproducible example to share? - July 21, 2016
- Few random questions on Random Forest - July 20, 2016